Sunteți pe pagina 1din 12

Structural Change and Economic Dynamics 44 (2018) 34–45

Contents lists available at ScienceDirect

Structural Change and Economic Dynamics


journal homepage: www.elsevier.com/locate/sced

Structural labour market reforms, GDP growth and the functional


distribution of income夽
Emiliano Brancaccio a,∗ , Nadia Garbellini b , Raffaele Giammetti c
a
Department of Law, Economics, Management and Quantitative Methods, University of Sannio, Via delle Puglie 82, Benevento, Italy
b
Department of Management, Information and Production Engineering, University of Bergamo, Viale Marconi 5, Dalmine, BG, Italy
c
Department of Economics and Social Sciences, Marche Polytechnic University, Piazzale Martelli 8, Ancona, Italy

a r t i c l e i n f o a b s t r a c t

Article history: The view that “structural reforms” designed to deregulate the labour market would be able to increase
Received 19 May 2017 employment and income is questioned by controversial empirical results, while the likelihood that these
Received in revised form 10 August 2017 reforms accentuate inequalities seems to be supported by recent evidence. The wide literature on this
Accepted 6 September 2017
issue, however, lacks studies on the specific implications of “structural labour-market reforms” on growth
Available online 20 October 2017
and functional distribution of income. The purpose of this study is to fill this gap by proposing two new
empirical analyses aimed at verifying the existence of statistical relationships between changes in the
JEL classifications:
Employment protection legislation index (EPL) on one side and variations of real GDP growth and wage
B5
C1
share on the other side. Results indicate that EPL reductions have no significant links with real GDP
E25 growth whereas they are significantly correlated with wage share reductions. These results contradict
some features of the neoclassical theories of growth and distribution.
Keywords: © 2017 Elsevier B.V. All rights reserved.
Structural reform
Labour market flexibility
EPL
GDP growth
Functional income distribution
Inequality
Neoclassical theory
Alternative approaches to growth and
distribution

1. Introduction pean countries in the sixties (Togliatti, 1964). From a theoretical


point of view, the modern concept of “structural labour-market
In the last quarter of a century, in many OECD countries and reforms” is neoclassical. In his definition of the “natural rate of
especially in Europe the prevailing economic policies have been unemployment”, Friedman (1968) refers to “structural character-
characterised by a marked tendency to implement “structural istics” of the labour market as the features that hinder efficiency:
labour-market reforms”. This widespread expression in the politi- market imperfections, stochastic variability in demands and sup-
cal debate does not have classical origins: it is not derived from the plies, the cost of gathering information about job vacancies and
Marxian concept of “structure” (Marx, 1867, 1885, 1894) or inspired labour availabilities, the costs of mobility, and so on. Therefore, the
by the more recent definitions of “structural dynamics” (Pasinetti, term “structural labour-market reforms”,is usually used to indicate
1993; Brian et al., 1991; Andreoni and Scazzieri, 2014), nor it has all measures aimed at removing the restrictions that interfere with
connections with the “structural reforms as a way to socialism”, the spontaneous operation of market forces in order to generate
an expression that pervaded the political debate in some Euro- a more efficient allocation of scarce resources and a consequent
growth of potential output (IMF, 2015; Canton et al., 2014; Turrini
et al., 2015). In this sense, the so-called “structural labour-market
夽 The authors would like to thank two anonymous referees for their valuable
reforms” consist in policies that reduce the amount of unemploy-
comments on previous drafts of this paper. The usual disclaimer applies.
ment benefits, remove the constraints to hiring and firing, decrease
∗ Corresponding author. the power of the trade unions, and so on, all in order to make the
E-mail addresses: emiliano.brancaccio@unisannio.it (E. Brancaccio), operation of the labour market equivalent to a “spot market for
nadia.garbellini@unipv.it (N. Garbellini), r.giammetti@pm.univpm.it a perishable commodity”, in the presumption that without con-
(R. Giammetti).

https://doi.org/10.1016/j.strueco.2017.09.001
0954-349X/© 2017 Elsevier B.V. All rights reserved.
E. Brancaccio et al. / Structural Change and Economic Dynamics 44 (2018) 34–45 35

straints such a market can ensure full and efficient employment of 1.80) and the variance decreased by 0.82 points (from 1.38 to 0.56).
labour and other production factors (Solow, 2004). Nowadays, the This path is in some ways even more accentuated in the aftermath of
meaning attributed to “structural reforms” seems to refer to the the so-called international “Great Recession” (IMF, 2012): the num-
views inspired by the neoclassical competitive general equilibrium, ber of labour market reforms has increased during the international
which under given assumptions concludes that all policies which economic crisis as these interventions have been viewed as impor-
favour the “spontaneous” play of market forces should lead to more tant policy tools to address growing unemployment (Adascalitei
efficient equilibrium levels of relative prices and of income distri- and Pignatti Morano, 2015). This decrease of the EPL was even more
bution corresponding to a full and better use of labour and other pronounced in the European Union, reflecting the political will to
scarce resources. According to this interpretation, a high degree of adapt to pursue the greater flexibility of the Anglo-American labour
labour market rigidity increase the bargaining power of workers, market (Berg, 2015). Between 1990 and 2013, within the first 11
and this would lead to a stickiness of wages and in general of the countries of the Euro Area − with the exception of Luxembourg
price system, which would be an impediment to the spontaneous because of lack of data − there were 40 negative changes of EPL,
balance of demand and supply of labour (Blanchard and Giavazzi, with a reduction of the average EPL index by 0.64 points (from 2.55
2003). Removal of such structural impediments, this theory con- to 1.91) and a decrease of its variance by 0.89 (from 1.19 to 0.3).
tends, would allow the achievement of the production frontier and Nevertheless, what were the real effects of “structural labour-
promote the economic growth. For example, in the context of the market reforms”? In contrast to the political sphere, nowadays
well-known Solow (1956) growth model, it can be assumed that at within the field of economic research there is no consensus on the
a certain level of worker protection correspond a real wage higher alleged macroeconomic benefits of deregulation policies. While in
than the one that guarantees the full utilization of the factors of the 90 s literature was inclined to be in favour of the flexibility of the
production. For a given stock of available capital, a limited num- labour market, with the accumulation of empirical evidence grow-
ber of workers will be assumed, corresponding to the point where ing doubts emerged. Skeptical positions increased about the actual
marginal productivity of labour coincides with the real wage fixed capacity of labour deregulations to reduce unemployment and
exogenously by labour protections. This leads to a high capital increase employment and income. As we shall see later, even insti-
labour ratio, which results in a low overall volume of savings and tutions such as the World Bank, the International Monetary Fund
therefore in a low capital accumulation and low income growth. and the OECD, which have repeatedly supported labour deregula-
Substantially similar conclusions also come to the other variants tion policies, today recognize that the empirical evidence available
of neoclassical growth theory: from the overlapping-generations does not confirm that such recipes can boost employment. The
models, to the intertemporal optimisation models, to models with same uncertainties about the effectiveness of deregulation in stim-
the classic hypothesis of saving, to the neoclassical versions of the ulating employment have fueled concerns about the effectiveness
new growth theory, for all these models, an economy characterised of these policies in reducing inequalities between so-called insid-
by rigid contracts and sticky wages grows at a lower rate than a ers and outsiders in the labour market. Moreover, in recent years, a
flexible economy (Schivardi, 1999). In particular, lack of labour flex- series of studies have highlighted that labour deregulation policies
ibility implies that the economy produces below its potential and can be related with inequality: Richard Freeman, among others,
slows convergence of the economy toward its steady state (Alonso suggested that the reforms aimed at deregulating labour market
et al., 2004); generates a redistribution of the product in favour of and making contracts more flexible generate tangible effects not so
the workers that compromises the process of capital accumulation much on employment and GDP, but rather on income inequalities.
(Bertola, 1994); and ultimately compresses profits and disincen- In this sense, as we shall see, some scholars are trying to provide
tives investment (Daveri and Tabellini, 2000). empirical evidence of what they call the ‘Freeman conjecture’.
The view that structural reforms would increase the efficiency In the literature, however, the analyses of the impact of labour
of the labour market, improve the competitiveness of the economy deregulation on employment and GDP growth are generally sepa-
and favour an increase in employment, production and income has rated from the studies on the link between deregulation and income
in the last twenty years prevailed among policymakers. Accord- distribution. In addition, existing analyses only look at the distribu-
ing to this interpretation, such reforms would even help to reduce tion between percentiles of population while not considering the
internal inequalities in the labour market, between insiders and classic problem of the functional distribution of income. There is
outsiders. After the Great Recession, the support in favour of the therefore a gap in literature: there are no joint studies devoted to
deregulation of the labour market has been enhanced by additional the relationship between structural reforms of the labour market on
elements, including the need to increase the flexibility of wages and the one hand and growth and distribution of income on the other.
prices in less competitive countries and contribute, in this way, to This research represents an attempt to fill this gap in the avail-
the absorption of the global macroeconomic imbalances. Impor- able literature. In this sense, we propose two novel analyses: a first
tant contributions to the achievement of this “deregulatory” vision test based on a balanced panel that investigates on whether or not a
of the labour market have been made by leaders of main govern- long-term relationship between deregulation on the one hand, and
ments and by the most influential international institutions (ECB, growth and functional distribution of income on the other, exists;
2011; EC, 2013; for a critical survey, see ILO, 2015). and a second test based on an unbalanced panel which examines
These guidelines in favour of “structural reforms” of the labour the possible relation between “shocks” in labour market regula-
market have had a significant impact on policies implemented by tion and trends of growth and functional distribution of income
national governments. Quantitative evidence of these reforms can in subsequent years. Both tests, as we shall see, appear consis-
be drawn from the ”Employment protection legislation” overall tent with the view that deregulation policies are associated with
index (EPL) calculated by the OECD. The EPL is an index of the changes in the functional income distribution in favour of profits
degree of regulation of the labour market: a decrease in EPL is usu- while seem not to be related with changes vin GDP growth.The
ally interpreted as a “deregulation” which reduces the “rigidities” choice of analysing the inequality in terms of functional distribu-
and increases the “flexibility” of the labour market (on this point, tion will allow us to discuss the empirical results also in the light of
see Howell et al., 2007; Skedinger, 2010; Adascalitei and Pignatti the classical debate between alternative theories of growth and dis-
Morano, 2015). The OECD reports historical series of EPL for 26 tribution. We will see that the results of the empirical tests contrast
countries. Between 1990 and 2013, 62 negative changes of EPL with the theoretical view, typically neoclassical, that relative prices
occurred in 16 of the 26 countries examined. Over the same period and functional distribution on the one hand and income growth on
the average of the EPL index decreased by 0.47 points (from 2.27 to the other are closely interlinked, so a policy shock that impacts on a
36 E. Brancaccio et al. / Structural Change and Economic Dynamics 44 (2018) 34–45

variable group should also have repercussions on the other. More in the empirical works of the 90 s and early 2000 s were largely biased
line with the empirical results of this study appear, instead, those by a methodological approach defined “theory-driven” (Howell
alternative theoretical approaches that exclude the possibility of et al., 2007): “While there are solid theoretical arguments under-
identifying univocal logical relationships between the functional pinning the call for such (labour) reforms, the empirical evidence
income distribution and GDP dynamics. is somewhat less developed and, in some cases, unsupportive”
The work is organized as follows. Section 2 is devoted to a review (IMF, 2003 p. 129). In fact, most recent analyses reveal the absence
of the literature on the relationship between “structural reforms” of of a clear relation between EPL and labour market performance.
the labour market on the one hand and employment, GDP growth, According to Olivier Blanchard, the IMF’s former chief economist:
wages and income distribution, on the other. Section 3 introduces “differences in employment protection seem largely unrelated to
the two empirical tests that we carried out in this paper. Section differences in unemployment rates across countries” (Blanchard,
4 describes the test based on a balanced panel, which analyses the 2006, p. 30). Others influential economists have faced this issue
long-running relationship between deregulation policies, on the with same results, for instance Baker et al. (2005, p. 109) states:
one hand, and growth and functional distribution, on the other. “the empirical case has not been made that could justify the
Section 5 describes the test based on an unbalanced panel built sweeping and unconditional prescriptions for labour market dereg-
in order to analyse the possible nexus with growth and functional ulation that pervade much of the policy discussion”, and Heckman
income distribution of any “shock” in labour regulation. Section 6 (2007, p. 4) commenting the evidence on the effects of regulation
examines the consistency of the empirical results obtained with the on European labour markets, admits: “the evidence currently in
neoclassical theory of growth and distribution of income and with play in this literature is weak” (similar findings are in Belot and
some alternative theories. Section 7 concludes. van Ours, 2004; Garibaldi and Violante, 2005; Baker et al., 2005;
Bassanini and Duval, 2006; Allard and Lindert, 2007; Amable et al.,
2007; Baccaro and Rei, 2007; Cazes and Nesporova, 2007; Griffith
2. Structural reforms, employment, GDP growth and et al., 2007; Rovelli and Bruno, 2008; Howell and Rehm, 2009;
income distribution: a short survey Armingeon and Baccaro, 2012; Storm and Naastepad, 2012; Vergeer
and Kleinknecht, 2012; Avdagic and Salardi, 2013; World Bank,
The consensus concerning the view that employment protection 2013; ILO, 2015; Adascalitei and Pignatti Morano, 2015).
regulations have a negative impact on macroeconomic outcomes, The new findings are confirmed in the recent statements of
causing an increase in unemployment, a decrease in productivity international organizations. According to a recent World Bank’s
and GDP growth and, moreover, generally worsening the efficiency Development Report: “New data and more rigorous methodolo-
and competitiveness of the labour market, has represented for long gies have spurred a wave of empirical studies over the past two
time the so-called “OECD-IMF orthodoxy” (Howell et al., 2007), decades on the effects of labour regulation. [. . .] Based on this
also known as “Transatlantic Consensus” (rephrased as “Berlin- wave of new research, the overall impact of EPL and minimum
Washington Consensus” in Fitoussi and Saraceno, 2013). This view, wages is smaller than the intensity of the debate would suggest.
established in the 90 s, took the flexibility of labour contracts in Most estimates of the impacts on employment levels tend to be
the United States as a benchmark and explained the increased insignificant or modest” (World Bank, 2013, p. 261). In a recent
unemployment in Europe as a result of excessive labour protec- IMF World Economic Outlook we can read: “The analysis shows
tions (Siebert, 1997; on this point see also Berg, 2015). According that reforms that ease dismissal regulations with respect to regu-
to the above interpretation, the 2003 IMF WEO (p. 129) states: “The lar workers do not have, on average, statistically significant effects
persistence of high unemployment in a number of industrial coun- on employment and other macroeconomic variables” (IMF, 2016,
tries – notably in continental Europe – is arguably one of the most p. 115). Similar conclusions can be found in the OECD Employment
striking economic policy failures of the last two decades. A wide Outlook: “Most empirical studies investigating medium/long-term
range of analysts and international organizations – including the effects of flexibility-enhancing EPL reforms suggests that they have,
European Commission, the OECD, and the IMF – have argued that at worst, no or a limited positive impact on employment levels in
the causes of high unemployment can be found in labour mar- the long run” (OECD, 2016, p. 126).
ket institutions. Accordingly, countries with high unemployment The economic research of the last few years comes to con-
have been repeatedly urged to undertake comprehensive struc- troversial results also with regard to the possible links between
tural reforms to reduce ‘labour market rigidities’ such as generous structural reforms in the labour market and the performances of
unemployment insurance schemes; high employment protection, GDP. Belot et al. (2007) studied the welfare effects of employment
such as high firing costs; high minimum wages; noncompetitive protection, concluding that there is a hump shaped relationship
wage-setting mechanisms; and severe tax distortions”. between EPL and growth, in the sense that an increase in EPL from
In fact, looking at the empirical literature, in the 90 s the analy- a low level leads to increased GDP per capita, but this positive
sis seem to suggest a convergence showing a positive relationship effect turns negative above a certain threshold. Similarly, Allard
between EPL and unemployment. One of the first studies on the and Lindert (2007) assumed the existence of a threshold presenting
impact of EPL on aggregate employment and unemployment is two sets of results on EPL, one for countries with little or no wage
Lazear (1990), which concludes by arguing that employment is coordination, and one for countries with closely coordinated wage
lower and unemployment is higher in countries with more strin- setting. In the first group, increased job protection did not seem to
gent EPL. The same conclusions are in the OECD’s Jobs Study (1994, reduce productivity and GDP growth, while, in contexts of strong
p. 28), that suggests: “An efficient and flexible supply side of the wage coordination, extra EPL lowers productivity and growth sig-
economy is crucial in ensuring that practices and policies oper- nificantly. With different models Lingens (2003) and Mortensen
ate in ways that create new jobs and help people fill them”. The (2005) found not conclusive results on the effects of labour dereg-
OECD remarks have had influence on the literature of the time (see ulation on economic growth. Lingens (2003) treats the impact of
Howell et al., 2007 and Berg, 2015). Indeed, in the late 90 s, three unions on GDP growth in a model with two kind of skills, show-
OECD economists found a significant and positive effect of EPL on ing ambiguous results. The same ambiguous results are shown in
unemployment (see Scarpetta, 1996 and Elmeskov et al., 1998). the matching model of schumpeterian growth built by Mortensen
However, over the last years some doubts emerged about the (2005). Adjemian et al. (2010) estimate a regression in which the
reliability of the first analyses on the relationship between EPL and endogenous variable is the average growth rate of GDP per capita of
employment. In particular, a view has emerged in the literature that each European Region and show a positive but not significant rela-
E. Brancaccio et al. / Structural Change and Economic Dynamics 44 (2018) 34–45 37

Table 1
Descriptive statistics, balanced panel.

Wage Share (WS) GDP growth (GDP g) EPL

Country Min Median Mean Max SD Min Median Mean Max SD Min Median Mean Max SD

All 43.2 57.0 56.6 72.9 5.3 −9.6 2.4 2.1 10.7 2.6 0.3 1.9 1.9 4.1 0.9
AUS 53.2 56.1 56.2 59.1 2.0 0.4 3.7 3.2 4.9 1.0 1.0 1.1 1.1 1.3 0.1
AUT 52.8 55.5 56.3 60.5 2.4 −3.9 2.2 1.9 3.6 1.6 1.8 2.0 1.9 2.0 0.1
BEL 58.4 61.0 60.8 62.9 1.2 −2.3 1.8 1.7 3.6 1.5 2.1 2.1 2.5 3.2 0.5
CAN 54.4 55.8 56.6 60.8 1.9 −2.7 2.6 2.3 5.0 1.9 0.6 0.6 0.6 0.6 0.0
CHE 62.7 65.3 65.1 67.6 1.3 −2.2 1.6 1.5 4.1 1.6 1.4 1.4 1.4 1.4 0.0
DEU 53.7 57.7 57.5 60.4 1.8 −5.8 1.7 1.3 4.0 2.1 1.8 2.3 2.3 3.0 0.5
DNK 54.0 55.7 55.9 59.3 1.3 −5.2 1.6 1.5 5.2 2.1 1.8 1.8 1.8 2.7 0.2
ESP 54.9 58.2 58.4 62.9 2.2 −3.6 2.7 1.9 5.2 2.5 2.3 2.8 2.9 3.6 0.4
FIN 52.2 55.2 55.9 65.5 3.7 −8.6 2.7 1.7 6.1 3.8 1.8 1.9 1.9 2.0 0.1
FRA 55.3 56.5 56.7 58.1 1.0 −3.0 1.9 1.6 3.8 1.5 2.7 3.0 3.0 3.0 0.1
GBR 55.1 58.6 58.3 60.4 1.5 −4.3 2.6 2.0 3.9 1.9 0.7 0.8 0.7 0.8 0.1
GRC 48.2 50.4 50.8 54.8 2.1 −9.6 2.8 0.9 5.6 4.3 2.2 3.8 3.3 3.8 0.6
IRL 44.7 49.3 50.7 59.3 4.9 −5.8 5.4 4.5 10.6 4.1 0.8 0.8 0.9 1.0 0.1
ITA 51.0 53.4 53.7 58.7 2.3 −5.6 1.5 0.7 3.6 2.0 2.3 2.8 3.0 3.8 0.6
JPN 58.7 63.0 63.0 67.3 3.1 −5.7 1.6 0.9 4.6 2.0 1.1 1.3 1.4 1.7 0.2
KOR 59.5 65.6 66.4 72.9 4.3 −5.6 5.3 5.2 10.7 3.5 2.3 2.3 2.5 3.1 0.4
NLD 56.1 59.8 60.0 64.3 2.3 −3.8 2.1 2.0 4.9 2.0 1.9 1.9 2.0 2.2 0.1
NOR 43.2 48.8 48.6 53.1 3.0 −1.6 2.6 2.4 5.1 1.6 2.5 2.7 2.7 2.8 0.1
NZL 47.7 51.3 51.0 54.1 1.7 −1.9 3.2 2.7 6.2 2.0 0.8 1.2 1.1 1.3 0.2
POL 47.6 54.9 53.8 63.0 5.1 −7.3 3.8 3.6 7.0 2.9 1.2 1.5 1.7 2.0 0.3
PRT 53.7 58.6 58.2 60.6 2.1 −4.1 1.8 1.3 4.7 2.4 2.5 3.7 3.6 4.1 0.4
SWE 46.5 48.9 48.8 51.8 1.4 −5.3 2.6 2.0 5.8 2.7 1.7 2.0 2.2 3.4 0.5
USA 56.8 59.7 59.5 61.5 1.5 −2.8 2.7 2.4 4.6 1.8 0.3 0.3 0.3 0.3 0.0

tion with the employment protection. Summarizing, in agreement with an increase in income inequality (Berg, 2015; ILO, 2015).
with the results regarding employment and unemployment, the With respect to the direct effect of employment protection laws on
empirical literature seems to question the possible impact of labour income inequality a research published by the IMF states: “there
market policy on the performances of GDP (for a wide review on is some evidence that lower employment protection is associated
EPL and its effects which confirms these conclusions, see Skedinger, with a higher Gini of gross income and that weaker employ-
2010). ment protection on temporary contracts is related to increased top
In the years in which the capacity of the structural reforms income shares—although the coefficient is small” (Jaumotte and
of labour market to stimulate employment and economic growth Osorio-Buitron, 2015 p. 21). Similarly, De et al. (2011), Darcillon
was questioned, there was also a growing interest in studies (2015), Calderón et al. (2005), Dosi et al. (2016), conclude that
devoted to the possible nexus between these reforms and inequal- stringent EPL is negatively associated with income inequality.
ity. According to Richard Freeman, more rigid labour regulations It is important to notice that the most of the studies on this sub-
“reduce the dispersion of earnings and income inequality”, while ject analyse the impact of labour deregulation on income inequality
they have “equivocal effects on other aggregate outcomes, such looking only at the Gini coefficient and the personal income distri-
as employment and unemployment”; therefore, “institutions alter bution. To the best of our knowledge, only a few recent studies
the distribution of income but not the efficiency of production” have analysed the inequality in terms of the functional distribu-
(Freeman, 2008 p. 15). This view has been named as the ‘Free- tion of aggregate income between profits and wages (Daudey and
man conjecture’ and tested in a comparative econometric analysis Garcia-Peñalosa, 2007; Atkinson, 2009; Stockhammer et al., 2008;
by Campos and Nugent (2015). According to their results, more Stockhammer, 2013; Jaumotte and Osorio-Buitron, 2015). Accord-
employment protections decrease inequality but have no signifi- ing to a recent IMF study: “Stronger labour market institutions
cant effect on GDP growth: then, institutional rigidities in labour could increase the wage share” (Dabla-Norris et al., 2015, p. 21).
market do not affect the aggregate income but can have an impact The empirical analysis in Guerriero and Sen (2012) supports this
on its distribution. This conclusion is supported by other contribu- view. These authors attempt to identify the driving force behind
tions which show a negative relationship between the strictness the decline of the labour share around the world. After an exten-
of employment law and income inequality: “Our results indicate sive econometric analysis focused primarily on the relationships
that labour market institutions exhibit significant correlations with between the labour share and measures of international trade and
the distribution of income across countries and over time. Stronger technological change, they conclude that regulations in the labour
institutions are correlated with lower inequality. [. . .] A central ele- market also matter. On the relationship between EPL and wage
ment of the so-called ‘flur result’ agreements that the EU proposes share, the OECD argues that: “Employment protection regulations
is a reduction in employment protection legislation. A consequence can affect the labour share. Indeed, stringent dismissal regulations
of such policy is likely to be a substantial increase in income might worsen the employer’s bargaining position”, so “(labour)
inequality” (Checchi and Garcia-Peñalosa, 2008, p. 633). On closer reform might have a negative impact on the labour share. However,
inspection, these results have also the endorsement of some of the there is little research on this issue” (OECD, 2012, p. 146).
most influential international organizations: “Many studies find Existing analyses, then, look at the income distribution between
that weakening labour market institutions can adversely impact percentiles of population while rarely considering the classic prob-
distribution of labour income shares at the bottom and middle of lem of the functional distribution of income between profits and
the distribution, through a variety of channels including by reduc- wages. Furthermore, as we have seen, the analyses of the impact of
ing workers’ bargaining power” (ILO-IMF-OECD-World Bank, 2015, labour deregulation on employment and GDP growth are usually
p. 21). Others studies point out that the emergence of non-regular separated from the studies on the link between deregulation and
contracts (part-time, temporary, etc) have also been associated income distribution. There is therefore a gap in literature: there
38 E. Brancaccio et al. / Structural Change and Economic Dynamics 44 (2018) 34–45

Fig. 1. Means and Confidence Intervals (95%), by country and year (1991–2013).

are no joint studies devoted to the relationship between structural which can hardly explain spurious effects captured by regression
reforms of the labour market on the one hand and growth and func- coefficients.
tional distribution of income on the other. Aiming to fill this gap, Then, we propose here two different approaches. The first one
in the next section we will present two empirical analyses aimed is in line with Campos and Nugent and is based on a set of OLS
at verifying the existence of statistical relationships between vari- regressions with fixed effects using a balanced panel. The second
ations of the Employment protection legislation index (EPL) on the approach, instead, is based on an additional set of estimates based
one hand and variations of the real GDP growth and the wage share on an unbalanced panel which focuses on the years before and after
on the other hand. the presence of relevant “shocks” in the level of EPL; this second
analysis represents an adaptation to labour market deregulations
of an original analysis of currency crises suggested by Maarek and
3. Two empirical exercises
Orgiazzi (2013) and further developed by Brancaccio and Garbellini
(2015). In what follows are descriptive statistics, methodological
In what follows, we are going to perform two different kinds
details and a description of the corresponding estimates which will
of empirical exercises in order to assess the effects of changes
be given separately for the two different exercises.
in the degree of labour protection on the rate of growth of GDP
and on functional income distribution. Campos and Nugent (2015)
provide a pioneering empirical support to the so-called ‘Freeman 4. Balanced panel
conjecture’, according to which labour market reforms do not affect
growth, but rather income distribution. We intend to test the same Table 1 provides some descriptive statistics of the balanced
conjecture by using a different methodological approach and alter- panel, which includes data for wage share, EPL, and the rate of
native data. growth of GDP for 23 countries, for the period 1991–2013. As we
The most relevant difference between the approach adopted can see, overall the wage share ranges from a minimum of 43.2 to
here and Campos and Nugent’s concerns the concept of income a maximum of 72.9, the standard deviation being 5.3. This wide
distribution: whereas they look at personal income distribution variability is mainly due to inter-country differences rather than
– described by the evolution of GINI coefficient – we are inter- to a high variability of wage shares through time. However, there
ested in functional income distribution, and hence we focus on are some countries in the dataset which show a relevant variabil-
the wage share. Secondly, Campos and Nugent decided to com- ity through time as well: Poland (from a minimum of 47.6 to a
pute themselves an index of Labour Market Legislation Rigidity maximum of 63.0), Ireland (44.7–59.3), Finland (52.2–65.5), Japan
(LAMRIG) based on ILO’s database of national labour, social secu- (58.7–67.3) and Norway (43.2–53.1).
rity and related human rights legislation. On the contrary, we made The rate of growth of GDP shows a wider gap between minimum
the choice here to stick to the more commonly used Employment and maximum levels, ranging from a minimum of −9.6% to a max-
Protection Legislation (EPL) index compiled by the OECD. LAMRIG imum of 10.7%, the standard deviation being 2.6. Also in this case,
covers the years from 1960s and a wider set of developing countries, of course, country-level data are characterised by a smaller gap
while EPL starts from 1985 and mainly concerns developed coun- between minima and maxima, with the exception of some coun-
tries. Our choice, then, reduces the number of countries included tries: Greece (minimum −9.6%, maximum 5.6%); Ireland (−5.8% to
in our dataset and the time span covered. However, it still permits 10.6%); Poland (−7.3% to 7.0%); Finland (-8.6% to 6.1%).
constructing a panel with 23 countries and 24 years. Considering As stated above, EPL shows basically no variability. Overall, it
the availability of data on wage shares – which are not published by ranges from a minimum of 0.3 to a maximum of 4.1, but at each sin-
the great majority of developing countries – using LAMRIG instead gle country level the gap between minima and maxima are much
of EPL would not have widened the extent of the present analysis. smaller, and standard deviation close to zero − showing that such
Thirdly, we adopted an alternative approach to overcome the main a gap is generated by a few once-and-for-all variations. Also in
problem emerging from using both EPL and LAMRIG for this kind of this case, there are some countries which show an above-average
analysis, i.e. the fact that these indexes exhibit a very low variability range of EPL levels: Germany (from 1.8 to 3, the standard devia-
E. Brancaccio et al. / Structural Change and Economic Dynamics 44 (2018) 34–45 39

Table 2 Table 3
Wage share, balanced panel. Rate of growth of GDP, balanced panel.

Dependent variable: Dependent variable:

WS GDP g

(1) (2) (3) (1) (2) (3)


(1991–2013) (1991–2007) (2008–2013) (1991–2013) (1991–2007) (2008–2013)

WS lag1 0.889*** 0.882*** 0.432*** GDP g lag1 0.374*** 0.306*** −0.123


(0.019) (0.023) (0.082) (0.035) (0.043) (0.084)
epl 0.449*** 0.496** 3.827*** epl 0.004 −0.005 0.047*
(0.162) (0.203) (1.116) (0.003) (0.003) (0.025)
d089 1.235*** d089 −0.037***
(0.170) (0.003)

Observations 525 387 138 Observations 505 367 138


R2 0.820 0.820 0.276 R2 0.396 0.134 0.038
Adjusted R2 0.780 0.767 0.226 Adjusted R2 0.376 0.124 0.031
F Statistic 759.528*** 821.979*** 21.524*** F Statistic 104.713*** 26.368*** 2.219 (df = 2;
(df = 3; 499) (df = 2; 362) (df = 2; 113) (df = 3; 479) (df = 2; 342) 113)

Note: *p < 0.1; **p < 0.05; ***p < 0.01. Note: *p < 0.1; **p < 0.05; ***p < 0.01.
Equation: WSi,t = ai + ␤1 WS lag1i,t + ␤2 epli,t + ␤3 d089 + ␧i,t. Equation: GDPgi,t = ai + ␥1 GDP g lag1i,t + ␥2 epli,t + ␥3 d089 + ␩i,t.

tion being 0.5); Spain (2.3–3.7); Greece (2.2–3.8); Italy (2.3–3.8);


Portugal (from 2.5 to 4.1). − which signals the response of income distribution to the sudden
Fig. 1 depicts means and 95% confidence intervals, by country drop in GDP − while the effect of the level of EPL is more than 8
and year, for EPL, WS and GDP g. As already emerging from Table 1, times the pre-crisis one. This means that during the crisis and the
EPL shows high inter-country variability, but is quite stable, though following recession, countries with a higher employment protec-
decreasing, through time. Moreover, inter-country variability is tion also kept a higher level of the wage share as against countries
itself decreasing through time, showing a tendency to convergence. with a more flexible labour market.
Wage share is also characterised by a huge heterogeneity among As to goodness of fit, the value of adjusted R squared shows
countries, but shows a tendency to decrease, with the exception of that estimates in columns (1) and (2) explain almost 80% of wage
2009 where average WS sharply increased due to the corresponding share variability between countries, while estimates in column (3)
drop of GDP growth induced by the crisis. only 22.6% – in other words, after the financial crisis a number of
The descriptive analysis has shown that, over the period consid- variables have been influencing the level of the wage share other
ered, on average both wage shares and EPL have been decreasing, than its historical trend and EPL.
while GDP growth does not show any clear tendency. This suggests, Table 3 shows the results of estimating Eq. (2). Also in this case,
in line with ‘Freeman conjecture’, the presence of some kind of cor- column (1) refers to the whole period, while columns (2) and (3) to
relation between EPL reductions − i.e., reductions of employment the pre- and post-crisis sub-periods, respectively. Looking at col-
protection − and redistribution of income from wages to profits. On umn (1), we can see that the level of employment protection had
the contrary, GDP growth shows no clear correlation with changes no influence on GDP growth. Only the lagged value of the rate of
in employment protection. growth of GDP itself and the crisis dummy had a significant effect,
In order to assess the relation between EPL and both the wage positive in the first case, negative in the latter. Again, estimates
share and the rate of growth of GDP, we estimated two OLS models in column (2) are qualitatively and quantitatively similar to those
with country-fixed effects, namely: in column (1), while column (3) shows a different picture. More
specifically, also in this case the coefficient associated to the lagged
WSi,t = ai + ˇ1 WS lag1i,t + ˇ2 epli,t + ˇ3 d089 + εi,t (1) value of the dependent variable becomes insignificant − showing
the structural break induced by the crisis − while the level of EPL
GDPgi,t = ai + 1 GDP g lag1i,t + 2 epli,t + 3 d089 + i,t (2)
becomes slightly positive and significant, showing that, during the
where ai are country dummies and d089 is a dummy taking value crisis and the following recession, countries with tighter employ-
1 for the years 2008 and 2009 − capturing the effect of the 2008 ment protection could limit the drop in GDP growth rates better
crisis. than countries with a more flexible labour market.
Tables 2 and 3 show the results of the estimation of Eqs. (1) and In this case, however, goodness of fit is definitely lower than in
(2) for the whole period – column (1) – and for the two sub-periods the previous one. Our estimates explain the 37.6% of GDP growth
1991–2007 and 2008–2013 – columns (2) and (3), respectively. rates variability for the whole period, 12.4% for the pre-crisis sub-
Looking at column (1) of Table 2, we can see a significant and period, and 0.31% only in the post-crisis sub-period. This means
positive effect of the lagged level of the wage share, the level of that the determinants of GDP growth are much more complex, and
EPL, and the crisis dummy. This means that higher EPL levels are that such a complexity significantly increased after the 2008 crisis.
significantly correlated to higher levels of the wage share. The pos- The conclusions which can be drawn from the empirical exercise
itive and significant effect of the crisis dummy indicates that, as performed in this section, therefore, are the following: (i) countries
we have already stressed in the descriptive analysis, wage shares with higher employment protection were on average characterised
were significantly higher in 2008 and 2009 with respect to the rest by higher levels of the wage share, both before and after the finan-
of the period considered here. In order to check for robustness of cial crisis; (ii) the insurgence of the crisis reinforced this relation,
the results, we performed a separate estimation for the two sub- making EPL an important element of wage share protection; (iii)
periods before (1991–2007) and after (2008–2013) the crisis itself. growth rates were, on average and over the period considered,
Results are shown in columns (2) and (3). The structural break is uncorrelated to the level of EPL; (iv) this relations appears as
apparent. While estimates in column (2) are almost the same as being slightly significant after the crisis only, suggesting that during
those in column (1), column (3) shows that after the crisis the effect recessions employment protection can be an important element in
of the lagged level of the wage share is half as it was before the crisis moderating GDP falls.
40 E. Brancaccio et al. / Structural Change and Economic Dynamics 44 (2018) 34–45

5. Unbalanced panel Therefore, the average value of negative percentage changes in


EPL (eplg ) has been corrected by a country-specific factor given by
Before describing the empirical strategy adopted here and the its coefficient of variation relative to that of the whole sample:
corresponding results, a methodological note is at place. Also in this    
case, as in Brancaccio and Garbellini (2015), no attempt is made at   CV eplg,i    eplg
˛i =  eplg,i   , CV eplg =   (3)
providing a causal interpretation to the statistical relations emerg- CV eplg  eplg
ing among the variables of interest. By adopting the distinction put
forward by Moneta and Russo (2014), we are working here with where i is a country-specific index. In other words, the threshold is
associational rather than causal models, the former being “statis- going to be higher for countries with above-average coefficient of
tical models in which associations among random variables are variation, and lower for countries with below-average coefficient
analysed”, and the latter “augmented statistical models in which of variation. Table 4 reports the values of the thresholds computed
some of the associations can be interpreted as causal relations” for each country on the basis of Eq. (3).
(Moneta and Russo, 2014, p. 55). Finally, we picked two values for exogenous thresholds: 2%
As stated in the introduction, the estimates presented in the pre- and 10%. Table 5 reports all the shocks which can be identified
vious section might be subject to some criticism. First and foremost, with the three alternative thresholds listed above. As can be seen,
EPL is almost stable through time, with just a few positive or neg- applying the endogenous threshold leads to the identification of
ative leaps. Moreover, the effects of a change in EPL may show up 29 shocks, 46 for the case of 2% and 23 for the case of 10% thresh-
with some delay. In order to confirm the results described above, we old. Table 6 provides some descriptive statistics for the unbalanced
estimated an alternative specification, aimed at overcoming these panels resulting from identifying EPL shocks by adopting the three
specific drawbacks. above-mentioned thresholds − the endogenous one, 2% and 10%,
More specifically, we picked the most significant leaps in the respectively − for the percentage change of EPL.
level of EPL, considering them as “shocks”, and concentrating atten- Looking at the wage share, we can see that in all three cases the
tion on the five years before and the five years after the shock minimum value increased from the pre-shock average to t+5, while
itself. After singling out all the “shocks”, estimates aim at checking the maximum value decreased, showing a polarisation towards
whether in any of t + i (i = 0, . . ., 5) there are significant differences, average values. Moreover, in all three cases both the median and
in the levels of wage share and rate of growth of GDP, with respect the mean have been decreasing over the time period considered. In
to pre-shock average. particular, looking at the mean, we observe a reduction in the wage
Before providing descriptive statistics of the unbalanced panel share in the year of the shock, followed by a slight recovery in t + 1
and describing the specification of the estimated equations and and t + 2 – with the exception of the 2% case, where recovery took
describing the corresponding results, it is worth spending a few place in t + 1 only – and then by a further reduction. Overall, with
lines to describe the way in which we defined shocks. As stated respect to the pre-shock average, the wage share decreased by 0.9
before, the methodology we are adopting here is based on Maarek p.p. in the first case (endogenous threshold), by 1.1 p.p. in the sec-
and Orgiazzi (2013), further developed by Brancaccio and Garbellini ond case (2%) and by 0.3 p.p. in the third case (shock corresponding
(2015). In those cases, however, the definition of “shocks” – or to an EPL reduction of at least 10%).
exchange rate crises – could take advantage of a well-established Turning to the rate of growth of GDP, Table 6 confirms the
literature and be based on the IMF currency regimes classification. absence of any well-defined trend, in all three cases, with min-
In this context, we do not have any supporting evidence or litera- imum, maximum, median and mean values oscillating over the
ture, and it is therefore necessary to look for a different solution. period considered.
A threshold can be defined either in an endogenous or exoge- Finally, the rate of change of EPL also shows a similar pattern in
nous, i.e. somehow arbitrary, way. Since both methods may be all three cases: with some exceptions, it is non-zero in the year of
subject to criticisms, we decided to take both, and hence to repeat the shock only. However, the minimum and maximum values show
estimates based on different definitions of shocks − a procedure that in some cases, the initial drop in EPL was followed by further
which also provides a check as to whether results are robust to drops, or by increases, of the index in the following years.
changes in the definition of shock itself. The descriptive analysis has shown that the wage share, in all
An endogenous definition relates the average value of the vari- three cases, underwent a reduction in the five years after the shock
able of interest, where an above-average value can be classified as with respect to the pre-shock average. Also here, as in the case of
abnormal. In this case, given the extremely heterogeneous char- the balanced panel, the rate of growth of GDP does not seem to be
acter of EPL percentage variations among different countries an affected by the EPL shock.
endogenous threshold should take this heterogeneity into account, In order to further investigate the relation between EPL shocks
and therefore be different from country to country. Looking at Fig. 2, and the wage share on the one side, and the rate of growth of GDP
we can see that there are countries, such as Belgium, where the EPL on the other side, we estimate two OLS models with country fixed
index is almost stable with only one single leap over the whole effects and a set of temporal dummies which allow to assess the
period. On the contrary, there are some countries, such as Portugal, time path followed by the variables of interest over the five years
where the index shows much more variability, the single changes after the shock. More specifically, the estimated equations are:
being less pronounced. The starting levels are also very different
from country to country, so that the same absolute change may 
5

5

WSi,t = ai + 1 WS lag1i,t + 2 d089 + j crt+j + ϑj crt+j flex (4)


translate into quite different percentage changes. A single thresh-
j=0 J=0
old, identifying the presence of a shock in Belgium, might fail to
consider any of Portuguese changes in EPL as abnormal, if too high,

5

5

or define almost all such changes as shocks, if too low. The more GDP gi,t = ai +
1 GDP g lag1i,t +
2 d089 + ıj crt+j + ϕj crt+j flex (5)
countries are present showing extreme changes in EPL, the more
j=0 J=0
simple average of negative shocks is inappropriate to represent an
endogenous threshold. where ai are country dummies, d089 is a dummy taking value
1 for the years 2008 and 2009, crt+j are the temporal dummies
taking value 1 j (j = 0,. . .,5) years after the shock, and flex is a dummy
taking value 1 when the rate of growth of EPL is negative, in order to
E. Brancaccio et al. / Structural Change and Economic Dynamics 44 (2018) 34–45 41

Fig. 2. EPL by country (1985–2013).

Table 4
Endogenous thresholds (˛i).

i Mean SD CV ˛i i Mean SD CV ˛i

All −9.7 0.06 −0.59 IRL −8.6 0.04 −0.51 −8.5


AUS −11.5 0.05 −0.45 −7.5 ITA −7.7 0.04 −0.58 −9.6
AUT −9.8 0.02 −0.19 −3.1 JPN −13.7 0.05 −0.35 −5.9
BEL −23.5 0.08 −0.35 −5.8 KOR −15.8 0.05 −0.30 −5.0
CZE −6.0 0.04 −0.64 −10.6 MEX −41.4 0.10 −0.24 −3.9
DEU −9.5 0.05 −0.49 −8.2 NLD −2.0 0.02 −1.00 −16.6
DNK −20.7 0.08 −0.37 −6.1 NOR −4.2 0.01 −0.36 −5.9
ESP −12.2 0.06 −0.45 −7.5 NZL −6.9 0.10 −1.41 −23.5
FIN −5.4 0.03 −0.47 −7.7 POL −18.4 0.11 −0.61 −10.2
FRA −3.0 0.02 −0.72 −12.0 PRT −6.5 0.03 −0.52 −8.6
GBR −10.3 0.03 −0.32 −5.4 SVK −15.8 0.11 −0.68 −11.3
GRC −11.1 0.07 −0.59 −9.9 SWE −11.6 0.06 −0.54 −8.9
HUN −13.5 0.05 −0.40 −6.7 TUR −1.1 0.00 −0.21 −3.5

capture the effects of further reductions in EPL after the one taking significance in all three cases, show some quantitative difference
place in time t. according to the adopted threshold.
As anticipated above, we adopted the same estimation strategy Looking first at column (1), showing the estimates correspond-
as in Brancaccio and Garbellini (2015), i.e. we first estimated the ing to the endogenous threshold, we can see that the initial shock
complete specification, and then we removed the regressor char- brought about a significant reduction of the wage share. The same
acterised by the lowest significance, up to the point where a further happens in all the five years after the shock, with the negative
removal would reduce the value of adjusted R squared. Results are effectincreasing in size up to t+2, and then decreasing in the follow-
shown in Tables 7 and 8. ing years, with the coefficient associated to crt+5 being only slightly
Table 7 shows the results of estimating Eq. (4) in correspondence smaller than that associated to crt (-0.482 versus −0.497).
between the three thresholds. In all three cases, the coefficient Looking at column (2), we can see that the significant and neg-
associated to the lagged value of the wage share is positive and sig- ative reduction brought about by the shock is smaller than in the
nificant, with almost the same magnitude − the highest value is the previous case (-0.444 versus −0.497), but is gradually increasing
one associated to the 10% threshold, followed by the one associated from t to t+5, with the only exception of a decrease in t+3 immedi-
to the endogenous threshold, and finally by that corresponding to ately followed by a further increase.
a 2% threshold. Looking at column (3), we can see that the time profile is similar
In the same way, the crisis dummy (d089) has a positive and to the one characterising the case of the endogenous threshold: a
significant effect in all three cases − the highest value being that negative and significant effect in time t is followed by increasingly
reported in column (1), followed by (2) and (3) − showing that in negative effects up to t+2, and then by decreasingly negative effects
2008 and 2009 the wage share on average increased with respect in the following year. The difference with respect to column (1) is
to the previous periods due to the sudden drop in GDP induced by that in time t+5 there is no significant change in the level of the
the crisis. wage share, showing that the effects of the shock were confined to
The estimates of the coefficients associated to the temporal the first 4 years after the shock itself.
dummies, though being qualitatively the same in terms of sign and In all three cases, the coefficients associated to the flex dummy
interacting with crt+j are not significant – and in almost all cases
42 E. Brancaccio et al. / Structural Change and Economic Dynamics 44 (2018) 34–45

Table 5
Availability, unbalanced panel.

Country Year Availability % Change Absolute Change Threshold

Endogenous 2% 10%

AUS 2007 2002–2012 −11.5 1.15 → 1.02 X X X


AUT 2003 1998–2008 −9.8 2.03 → 1.84 X X
BEL 1998 1993–2003 −42.7 3.24 → 2.12 X X X
BEL 2012 2007–2013 −4.4 2.23 → 2.13 X
CZE 2007 2002–2012 −6.0 2.22 → 2.09 X
DEU 1995 1991–2000 −2.0 2.96 → 2.90 X
DEU 1997 1992–2002 −11.5 2.90 → 2.59 X X X
DEU 2003 1998–2008 −11.3 2.34 → 2.09 X X X
DNK 1991 1986–1996 −40.0 2.65 → 1.78 X X X
ESP 1995 1990–2000 −26.3 3.65 → 2.81 X X X
ESP 2007 2002–2012 −4.6 2.81 → 2.68 X
ESP 2011 2006–2013 −11.7 2.68 → 2.38 X X X
FIN 1992 1987–1997 −8.8 2.02 → 1.85 X X
FIN 1997 1992–2002 −3.9 1.85 → 1.78 X
FIN 2002 1997–2007 −3.7 1.94 → 1.86 X
FRA 1987 1985–1992 −4.5 2.83 → 2.70 X
GBR 2013 2008–2013 −10.3 0.82 → 0.74 X X X
GRC 2004 1999–2009 −30.8 3.77 → 2.77 X X X
GRC 2011 2006–2013 −17.3 2.77 → 2.34 X X X
HUN 2013 2008–2013 −13.5 1.62 → 1.42 X X X
IRL 2006 2001–2011 −8.6 1.03 → 0.95 X X
ITA 1988 1985–1993 −4.7 4.00 → 3.82 X
ITA 1998 1993–2003 −16.2 3.75 → 3.19 X X X
ITA 2000 1995–2005 −6.1 3.19 → 3.00 X
ITA 2002 1997–2007 −15.6 3.00 → 2.57 X X X
JPN 1997 1992–2002 −5.8 1.70 → 1.60 X
JPN 2000 1995–2005 −21.5 1.60 → 1.29 X X X
JPN 2007 2002–2012 −13.7 1.29 → 1.12 X X X
KOR 1998 1993–2003 −11.5 3.08 → 2.75 X X X
MEX 2013 2008–2013 −41.4 3.10 → 2.04 X X X
NLD 1994 1989–1999 −3.9 2.23 → 2.14 X
NLD 1999 1994–2004 −10.0 2.11 → 1.91 X
NOR 2000 1995–2005 −3.5 2.76 → 2.66 X
POL 2003 1998–2008 −18.4 1.49 → 1.24 X X X
PRT 1990 1985–1995 −2.0 4.19 → 4.11 X
PRT 1992 1987–1997 −3.1 4.11 → 3.98 X
PRT 1997 1992–2002 −7.4 3.98 → 3.69 X
PRT 2004 1999–2009 −5.7 3.69 → 3.49 X
PRT 2008 2003–2013 −9.3 3.49 → 3.18 X X
PRT 2010 2005–2013 −4.7 3.18 → 3.04 X
PRT 2012 2007–2013 −9.9 3.04 → 2.75 X

Table 6
Descriptive statistics, unbalanced panel

Wage Share Rate of growth of GDP Rate of growth of EPL

Period Min Median Mean Max SD Min Median Mean. Max. SD Min Median Mean Max SD

Endogenous Threshold
Before 37.2 56.0 54.7 72.9 7.2 −6.8 2.0 1.9 10.3 2.8 −21.1 0.0 −0.5 30.1 4.7
t 37.1 54.0 53.8 69.1 7.0 −9.6 1.6 0.9 6.1 3.3 −42.7 −15.6 −18.0 −8.6 9.9
t+1 42.3 53.8 54.0 66.0 5.9 −7.6 1.5 1.2 10.7 3.8 −20.1 0.0 −0.3 47.3 11.2
t+2 41.8 53.6 54.0 65.0 5.5 −5.7 2.0 2.2 8.5 3.3 −19.8 0.0 −1.6 0.0 4.3
t+3 41.4 53.5 53.8 65.6 5.6 −5.8 3.3 3.0 10.3 3.0 −9.9 0.0 0.6 21.8 5.2
t+4 41.9 53.2 53.6 64.7 5.5 −4.1 2.2 2.3 7.2 2.6 −15.6 0.0 0.1 30.1 7.7
t+5 45.1 53.9 53.8 65.1 5.1 −5.6 1.6 1.5 6.1 2.7 −9.9 0.0 −1.5 0.0 3.2

Threshold 2%
Before 37.2 56.8 55.8 72.9 6.6 −6.8 2.5 2.3 10.3 2.6 −21.1 0.0 −0.3 30.1 4.2
t 37.1 56.2 55.0 69.1 6.5 −9.6 2.0 2.0 7.6 3.0 −42.7 −9.9 −12.5 −2.0 10.1
t+1 41.6 56.7 55.3 66.0 5.8 −7.6 1.8 1.5 10.7 3.2 −25.8 0.0 −0.5 47.3 9.5
t+2 42.3 56.1 55.1 65.3 5.2 −5.7 2.0 1.8 8.5 3.1 −19.8 0.0 −1.9 0.0 4.5
t+3 41.8 55.0 55.0 65.6 5.2 −5.8 2.6 2.4 8.1 2.4 −21.5 0.0 −0.8 21.8 5.8
t+4 41.4 54.4 54.8 64.7 5.3 −4.1 1.9 2.2 10.3 2.5 −15.6 0.0 −1.0 10.0 4.1
t+5 41.9 54.5 54.7 65.1 5.2 −4.4 1.7 1.7 6.1 2.4 −10.0 0.0 −0.6 30.1 6.0

Threshold 10%
Before 37.2 55.5 54.7 72.9 7.6 −6.8 2.0 1.9 10.3 2.8 −21.1 0.0 −0.3 30.1 4.0
t 37.1 53.9 53.7 69.1 7.3 −9.6 1.8 1.0 5.1 3.3 −42.7 −17.3 −20.3 −10.3 9.9
t+1 42.3 53.5 54.2 66.0 6.2 −7.6 1.5 1.2 10.7 4.0 −20.1 0.0 0.1 47.3 12.7
t+2 41.8 53.7 54.2 65.0 5.9 −5.7 2.0 2.3 8.5 3.6 −19.8 0.0 −1.7 0.0 4.8
t+3 41.4 53.7 54.0 65.6 6.2 0.8 3.2 3.6 10.3 2.3 −9.9 0.0 0.8 21.8 5.9
t+4 41.9 53.7 53.9 64.7 6.1 −0.5 2.3 2.7 7.2 2.4 −15.6 0.0 0.8 30.1 8.4
t+5 45.1 54.2 54.4 65.1 5.4 −5.6 1.5 1.3 5.2 2.8 −9.9 0.0 −1.1 0.0 2.9
E. Brancaccio et al. / Structural Change and Economic Dynamics 44 (2018) 34–45 43

Table 7 were even dropped by the estimated equations due to their very
Wage share, unbalanced panel.
low significance – showing that the fact of having further EPL
Dependent variable: reductions or not after the initial shock did not change the overall
performance of the wage share in the following five years.
WS
As to the goodness of fit, the highest value of adjusted R squared
(1) (2) (3) is the one in column (3) – which is of course associated to the
Threshold Endogenous 2% 10%
highest number of observations – followed by columns (1) and (2).
WS lag1 0.775*** 0.753*** 0.789*** Table 8 shows the results of estimating Eq. (5). In all three cases,
(0.039) (0.030) (0.041)
there is a positive and significant effect of the lagged value of the
d089 1.542*** 1.325*** 1.270***
(0.211) (0.191) (0.225) rate of growth of GDP, the strongest being associated to the 2%
crt −0.497** −0.444*** −0.515** threshold, the weakest to the 10% one. The crisis dummy is asso-
(0.207) (0.166) (0.214) ciated to a significant and negative coefficient, with no relevant
crt+1 −0.741*** −0.491*** −0.624*** difference between the three cases. We observe a significant and
(0.222) (0.174) (0.230)
crt+2 −0.747*** −0.494*** −0.803***
negative coefficient associated to crt in columns (1) and (3), though
(0.235) (0.183) (0.239) significance is at 95% only, and the value of the coefficient is close to
crt+3 −0.648*** −0.404** −0.641** zero (-0.008 and −0.010, respectively). All the remaining estimated
(0.248) (0.186) (0.258) coefficients are statistically non significant.
crt+4 −0.662*** −0.533*** −0.499*
Even in this case, as for the balanced panel, goodness of fit is
(0.253) (0.190) (0.262)
crt+5 −0.482* −0.570*** lower for GDP growth than for the wage share.
(0.258) (0.194)
crt+2cont −1.792
(1.104) 6. A theoretical interpretation of the empirical findings
Observations 290 465 228
R2 0.732 0.697 0.740 The empirical findings of this study raise a theoretical problem.
Adjusted R2 0.639 0.614 0.642
Not all existing paradigms can admit that a particular economic
F Statistic 86.518∗∗∗ (8;253) 104.742∗∗∗ 80.309∗∗∗
df = 8; 253 df = 9; 410 df = 7; 198
policy strategy has implications for functional income distribution
but not on income growth. The fact that “structural labour-market
Note: ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗ p < 0.01.
reforms” present significant statistical relationships with wage

5

5

Equation: WSi,t = ai + 1 WS lag1i,t + 2 d089 + j crt+j + ϑj crt+j flex. shares but not with real GDP growth rates is a difficult problem
to solve in the context of those analytical approaches that consider
j=0 J=0
these variables irreducibly interconnected. Let us consider the var-
ious versions of the neoclassical theory. In the basic concepts of
the temporary or intertemporal general competitive equilibrium
(Hicks, 1939; Arrow and Debreu, 1954), as well as in the stan-
Table 8 dard growth models (Solow, 1956; Koopmans, 1965 and others), an
Rate of growth of GDP, unbalanced panel.
exogenous shock that affects the relative prices and the distribu-
Dependent variable: tion of income between wages and profits must have an impact on
GDP g the level and growth of production, and vice versa. This also holds
for modern “imperfectionist” versions of the neoclassical theory.
(1) (2) (3)
Even if market imperfections, asymmetries of information, strate-
Threshold Endogenous 2% 10%
gic interactions and social institutions are admitted, these models
GDP g lag1 0.367*** 0.406*** 0.307*** preserve in terms more or less explicit the idea of a strict intercon-
(0.058) (0.043) (0.069)
d089 −0.039*** −0.036*** −0.039***
nection between functional distribution and growth of aggregate
(0.005) (0.004) (0.006) income (see Diamond, 1984; Williamson, 1986; Boyd and Prescott,
crt −0.008* −0.002 −0.010* 1986; Azariadis and Smith, 1996, among many others; for discus-
(0.005) (0.003) (0.005) sion on the term “imperfectionist”, see Eatwell and Milgate, 2011).
crt+1 −0.002 −0.006 −0.005
The reason is that neoclassical theory interprets the relative prices
(0.006) (0.004) (0.007)
crt+2 0.008 0.007 and the distribution variables as indicators of the relative scarcity of
(0.005) (0.006) factors of production, which ensure the equilibrium of their supply
crt+3 0.006 0.009 with the respective demand. Therefore, in all variants of neoclassi-
(0.005) (0.006) cal general equilibrium analysis any shock that changes the efficient
crt+4 −0.003
(0.004)
proportions of factor endowments will inevitably involve changes
crt+5 −0.0004 −0.003 −0.004 in both the functional income distribution and the income growth
(0.005) (0.004) (0.006) achieved with those factors proportions. In particular, a “structural
crt+2cont 0.013 0.012 0.015 labour-market reform” that leads to a new efficient allocation of
(0.012) (0.009) (0.013)
labour endowment compared to other existing factors, will result
Observations 288 459 226 in a corresponding change in both the wage share and the level and
R2 0.290 0.293 0.269 growth of real GDP. When analysing the whole economic system by
Adjusted R2 0.253 0.259 0.232
a neoclassical point of view, the determination of the social product
F Statistic 12.835∗∗∗ (8;253) 24.031∗∗∗ 8.966∗∗∗
df = 8; 251 df = 7; 406 df = 8; 195 and its growth must be simultaneous and symmetrical with respect
to the determination of the real wage and other distributive vari-
Note: ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗ p < 0.01.
ables. The likelihood that an exogenous policy change only affects

5

5

Equation: GDP gi,t = ai +


1 GDP g lag1i,t +
2 d089 + ıj crt+j + ϕj crt+j flex. one of the two groups of variables leaving unchanged the other, in
general, must be excluded. This aspect makes the neoclassical the-
j=0 J=0
ory and its modern variants unlikely compatible with the empirical
analysis presented in this paper (among many other examples of
44 E. Brancaccio et al. / Structural Change and Economic Dynamics 44 (2018) 34–45

contrast between modern variants of neoclassical theory and data, to verify whether and to what extent it could be conditioned by
see Stirati 2016). elements not covered in the first instance.
The empirical results of this study, as well as the lack of empirical
evidence about the neoclassical claim that labour-market reforms 7. Conclusions
reduce unemployment, seem to find an easier interpretation in light
of those alternative approaches that, rejecting the neoclassical view The two empirical analyses presented in this study indicate
of prices and distributive variables such as indicators of the relative that the so-called “structural reforms” of the labour market imple-
scarcity of factors, do not contemplate strict logical relationships mented in many OECD countries between 1990 and 2013 are
between growth and functional income distribution. A basic ref- statistically associated with divergences in functional income dis-
erence for this different approach can be drawn from the Ricardo’s tribution rather than GDP growth. From the balanced panel we find
method of analysis (Ricardo, 1817), then developed by Sraffa (1960) that a reduction in the level of labour protection (EPL) has no sta-
and his successors. This methodology can be explained in terms tistically significant positive impact on the growth of GDP whereas
of analysis that includes two different phases: the first stage is it is on average associated with a wage share about 0.5 percentage
dedicated to the determination of relative prices and income dis- points lower. Moreover, the unbalanced panel shows that shocks
tribution by assuming that the quantities produced, its technology in the labour protections corresponding to a decrease in EPL index
and a distributive variable can be considered exogenous, while a between 0.4 and 0.5 points are associated, in the five years follow-
second stage is dedicated to the analysis of the various, possible ing, with a cumulative reduction of the wage share between 3 and
determinants of those exogenous and to the eventual links exist- 4 percentage points. These results seem to be hardly conceivable in
ing between them (Garegnani, 1984; Roncaglia, 1978; Palumbo, the framework of neoclassical theory and its more recent variants,
2015). The first stage relates to some general and fundamental where any “structural reform” which influences functional income
characteristics of the economic system, whereas the second stage distribution should also have a simultaneous impact on income
is dedicated to the investigation of those elements of the system growth, and viceversa.
dependent on the institutional framework and, at the limit, easily
influenced by contingent factors. This separation of the analysis in
two logical stages contrasts with the methodology of neoclassical References
theory. These theory, in fact, relies on the spontaneous demand- Adascalitei, D., Pignatti Morano, C., 2015. Labour Market Reforms Since the Crisis:
supply mechanism to simultaneously determine the distribution Drivers and Consequences. Research Department Working Paper 5.
variables and the volume and composition of the product. On the International Labour Organization, Geneva.
Adjemian, S., Langor, F., Quintero-Rojas, C., 2010. How do labor market institutions
contrary, within the classical analysis, the division of the product
affect the link between growth and unemployment: the case of the European
between capital and labour does not depend on the interaction countries. Eur. J. Comp. Econ. 7 (2), 347–371.
between supply and demand. The distribution is governed by his- Allard, G.J., Lindert, P.H., 2007. EuroProductivity and Euro-Jobs since the 1960:
torical and social factors. In light of this approach, the empirical which institutions really mattered? In: Hatton, T.J., O’Rourke, K.H., Taylor, A.M.
(Eds.), The New Comparative Economic History: Essays in Honor of Jeffrey G.
results that show absence of relationships between EPL and unem- Williamson. MIT Press, Cambridge, MA, US, and London, UK.
ployment and EPL and growth, are not surprising. In fact, within the Alonso, A., Echeverria, C., Tran, K.C., 2004. Long-run economic performance and the
classicalframework, the employment protection legislation index labor market. South. Econ. J. 70 (4), 905–919.
Amable, B., Demmou, L., Gatti, D., 2007. Employment Performance and
is one of the elements of the historical and social factors which Institutions: New Answers to an Old Question. Discussion Paper No 2731 IZA,
determine relative prices and income distribution between social Bonn, Germany.
classes, in preliminary theoretical phase compared to that which Andreoni, A., Scazzieri, R., 2014. Triggers of change: structural trajectories and
production dynamics. Camb. J. Econ. vol. 38 (6), 1391–1408, Oxford University
determines the level and growth of product and income. Against Press.
the neoclassical method of simultaneous determinations, we can Armingeon, K., Baccaro, L., 2012. Do Labor Market Liberalization Reforms Pay Off?
therefore speak of a Ricardian approach of “causal” or “decom- University of Geneva, Mimeo.
Arrow, K.J., Debreu, G., 1954. Existence of an equilibrium for a competitive
posable” type, that defines the prices and distribution outside the
economy. Econometrica 22 (3), 265–290.
production system and before it, and which finds similarities even Atkinson, A.B., 2009. Factor shares: the principal problem of political economy?
with some Keynes’s logical arguments (Pasinetti, 1974). An appli- Oxf. Rev. Econ. Policy 25, 3–16.
Avdagic, S., Salardi, P., 2013. Tenuous link: labour market institutions and
cation of this approach is to be found in those studies that have
unemployment in advanced and new market economies. Soc. Econ. Rev. 11 (4),
suggested a compatibility between the Ricardian determination of 739–769.
prices and functional distribution, on the one side, and the deter- Azariadis, C., Smith, B.D., 1996. Private information, money, and growth
mination of the level and growth of income based on the Keynesian indeterminacy, fluctuations, and the mundell-tobin effect. J. Econ. Growth 1,
309–332.
principle of effective demand, on the other side (Garegnani, 1992; Baccaro, L., Rei, D., 2007. Institutional determinants of unemployment in OECD
Eatwell and Milgate, 2011). Under this alternative view, the pos- countries: does the deregulatory view hold water? Int. Organ. 61, 527–569.
sibility of finding linkages between the functional distribution of Baker, D., Glyn, A., Howell, D., Schmitt, J., 2005. Labor market institutions and
unemployment: a critical assessment of the cross- country evidence. In:
income and the growth of income and employment is by no means Howell, D. (Ed.), Questioning Liberalization: Unemployment, Labor Markets
excluded (see, for example, Kurose, 2013). However, the existence and the Welfare State. Oxford University Press, Oxford, pp. 72–118.
of rigid and unequivocal connections between these variables is Bassanini, A., Duval, R., 2006. The determinants of unemployment across OECD
countries: reassessing the role of policies and institutions. OECD Econ. Stud.
not an indispensable element of this approach. Therefore, within 42, 7–86.
this alternative paradigm, the existence of a statistically signifi- Belot, M., van Ours, J., 2004. Does the recent success of some OECD countries in
cant relationship between the so-called “structural reforms” aimed lowering their unemployment rates lie in the clever design of their labor
market reforms? Oxf. Econ. Pap. 56, 621–642.
at the deregulation of the labour market on the one side, and the
Belot, M., Boone, J., van Ours, J., 2007. Welfare-Improving employment protection.
reduction of the wage share on the other side, can be regarded as Economica 74, 381–396.
an empirical result that relates to the determination of the system Berg, J. 2015. Labour markets, institutions, and inequality, ed. Edward Elgar and
ILO, Cheltenham and Geneva.
of relative prices and income distribution among social classes. The
Bertola, G., 1994. Flexibility, investment and growth. J. Monetary Econ. 34 (2),
fact that this result is joined with the absence of a significant statis- 215–238.
tical relationship between the “structural labour-market reforms” Blanchard, O., Giavazzi, F., 2003. Macroeconomic effects of regulation and
and the growth of GDP can instead be regarded as an empirical deregulation In goods and labor markets. Q. J. Econ. 118 (August (3)), 879–907
(MIT Press).
result relating to the second stage of the analysis, not necessarily Blanchard, O., 2006. European unemployment: the evolution of facts and ideas.
general, that must be subjected to further investigation in order Econ. Policy 45, 5–47.
E. Brancaccio et al. / Structural Change and Economic Dynamics 44 (2018) 34–45 45

Boyd, J., Prescott, E., 1986. Financial intermediary-coalitions. J. Econ. Theory 38 (2), and Employment Ministers Meeting and Joint Meeting with the G20 Finance
221–232. Ministers, Ankara, Turkey, 3–4 September 2015.
Brancaccio, E., Garbellini, N., 2015. Currency regime crises, real wages, functional IMF, 2003. Unemployment and Labour Market Institutions: Why Reforms Pay Off.
income distribution and production. Eur. J. Econ. Econ. Policies 12 (3), 255–276. In World Economic Outlook 2003. IMF, Washington, DC, pp. 129–150.
Brian, A.W., Landesmann, M., Scazzieri, R., 1991. Dynamics and Structures. IMF, 2015. Structural Reforms And Macroeconomic Performance: Initial
Structural Change and Economic Dynamics, vol. 2., June (1) Elsevier, pp. 1–7. Considerations For The Fund, IMF Staff Reports, November 2015. IMF,
Calderón, C., Chong, A., Váldes, R., 2005. Labor market regulations and income Washington, DC.
inequality: evidence for a panel of countries. In: Restrepo, J.E., Tokman, A.R. IMF, 2016. Time for a Supply Side Boost? Macroeconomic Effects of Labor and
(Eds.), Labor Markets and Institutions. Central Bank of Chile, Santiago, Chile. Product Market Reforms in Advanced Economies. In World Economic Outlook
Campos, N.F., Nugent, J.B., 2015. The Freeman Conjecture. IZA/World Bank 2016. IMF, Washington, DC, pp. 101–142.
Conference on Employment and Development: Technological Change and Jobs, Jaumotte, F., Osorio-Buitron, C., 2015. Inequality and Labor Market Institutions.
Bonn. IMF Staff Discussion Note. IMF, Washington, DC.
Canton, E., Grilo, I., Monteagudo, J., Pierini, F., Turrini, A., 2014. The Role of Koopmans, T.C., 1965. On the Concept of Optimal Economic Growth. The Economic
Structural Reform for Adjustment and Growth. DG ECFIN Economic Brief, Approach to Development Planning. Rand McNally, Chicago, pp. 225–287.
Brussels. Kurose, K., 2013. The dynamics of the labour market and income distribution in
Cazes, S., Nesporova, A., 2007. Labour markets in central and south-Eastern europe: relation to the speed of demand saturation. Struct. Change Econ. Dyn. 24,
from transition to stabilization. In: Cazes, S., Nesporova, A. (Eds.), Flexicurity: A 101–111.
Relevant Approach in Central and Eastern Europe. ILO, Geneva, Switzerland. Lazear, E., 1990. Job security provisions and employment. Q. J. Econ. 105, 699–726.
Checchi, D., García-Peñalosa, C., 2008. Labour market institutions and income Lingens, J., 2003. The impact of a unionised labor market in a Schumpeterian
inequality. Econ. Policy 56, 601–649. growth model. Labour Econ. 10, 91–104.
Dabla-Norris, E., Kochhar, K., Suphaphiphat, N., Ricka, F., Tsounta, E., 2015. Causes Maarek, P., Orgiazzi, E., 2013. Currency crises and the labour share. Economica 80
and Consequences of Income Inequality: A Global Perspective. IMF Staff (319), 566–588.
Discussion Note. IMF, Washington, DC. Marx, K.,1867, 1885, 1894. Das Kapital, ed. Meissner, Hamburg; it. trans., Il Capitale,
Darcillon, T., 2015. How does finance affect labor market institutions? An empirical Cantimori, D., Panzieri, R., Boggeri, M.L. (Eds.), Editori Riuniti, Roma, 1994.
analysis in 16 OECD countries. Soc. Econ. Rev. 13 (3), 477–504. Moneta, A., Russo, F., 2014. Causal models and evidential pluralism in
Daudey, E., Garcia-Peñalosa, C., 2007. The personal and the factor distribution of econometrics. J. Econ. Methodol. 21 (1), 54–76.
income in a cross-section of countries. J. Dev. Stud. 43, 812–829. Mortensen, D.T., 2005. Growth, unemployment and labor market policy. J. Eur.
Daveri, F., Tabellini, G., 2000. Unemployment, growth and taxation in industrial Econ. Assoc.
countries. Econ. Policy Eur. Forum 30, 47–88. OECD, 1994. OECD Jobs Study. OECD, Paris.
De, S., Singh, M., Subramanian, K., 2011. Income inequality and employment OECD, 2012. Labour Losing to Capital: What Explains the Declining Labour Share?
protection laws: evidence from OECD countries. In: 7th Annual Conference on In OECD Employment Outlook 2012. OECD, Paris.
Economic Growth and Development, ISI Delhi, December, 2011. OECD, 2016. Short-term Labour Market Effects of Structural Reforms: Pain Before
Diamond, D., 1984. Financial intermediation and delegated monitoring. Rev. Econ. the Gain? In OECD Employment Outlook 2016. OECD, Paris.
Stud. 51, 393–414. Palumbo, A., 2015. Demand and supply. In: Kurz, H., Salvadori, N. (Eds.), The Elgar
Dosi, G., Pereira, M.C., Roventini, A., Virgillito, M.E., 2016. The Effects of Labour Companion to David Ricardo. Cheltenham, Edward Elgar, pp. 105–114.
Market Reforms upon Unemployment and Income Inequalities: an Agent Pasinetti, L.L., 1974. Growth and Income Distribution: Essays in Economic Theory.
Based Model. LEM Papers Series 2016/27. Laboratory of Economics and Cambridge University Press, Cambridge.
Management (LEM), Sant’Anna School of Advanced Studies, Pisa, Italy. Pasinetti, L.L., 1993. Structural Economic Dynamics –A Theory of the Economic
EC, 2013. Statement by president barroso on the country-specific Consequences of Human Learning. Cambridge University Press Cambridge, UK.
recommendations package 2013. In: European Commision Press Conference, Ricardo, D., 1817. On the Principles of Political Economy and Taxation, ed. John
Brussels, 29 May 2013 http://europa.eu/rapid/press-release SPEECH-13-473 Murray, London; it. trans. Principi di economia politica e dell’imposta. UTET,
en.htm/. Torino.
ECB, 2011. Introductory statement to the press conference. In: Mario Draghi Roncaglia, A., 1978. Sraffa and the Theory of Prices. John Wiley and Sons, New York.
President of the ECB, European Central Bank, Frankfurt am Main, 8 December Rovelli, R., Bruno, R., 2008. Labor Market Policies, Institutions and Employment
2011 https://www.ecb.europa.eu/press/pressconf/2016/html/is161208. en. Rates in the EU-27. Discussion Paper No 3502. IZA, Bonn, Germany.
html/. Scarpetta, S., 1996. Assessing the role of labour market policies and institutional
Eatwell, J., Milgate, M., 2011. The Fall and Rise of Keynesian Economics. Oxford settings on unemployment: a cross-Country study. OECD Econ. Stud. 26, 43–98.
University Press. Schivardi, F., 1999. Rigidità nel mercato del lavoro, disoccupazione e crescita, Temi
Elmeskov, J., Martin, J.P., Scarpetta, S., 1998. Key lessons for labour market reforms: di discussione, Banca D’Italia, Servizio Studi.
evidence from OECD countries’ experiences. Swedish Econ. Policy Rev. 5, Siebert, H., 1997. Labor market rigidities: at the root of unemployment in Europe. J.
205–252. Econ. Perspect. 11 (3), 37–54.
Fitoussi, J.-P., Saraceno, F., 2013. European economic governance: the Skedinger, P., 2010. Employment Protection Legislation. Evolution, Effects,
Berlin–Washington consensus. Camb. J. Econ. 37 (3), 479–496. Winners and Losers. Edward Elgar.
Freeman, R., 2008. Labor Market Institutions Around the World. London,LSE CEP Solow, R.M., 1956. A contribution to the theory of economic growth. Q. J. Econ. 70
Discussion Paper No. 844. (1), 65–94.
Friedman, Milton, 1968. The role of monetary policy. Am. Econ. Rev. 58 (1), 1–17. Solow, R.M., 2004. Introduction. In: Solow, R. (Ed.), Structural Reform and
Garegnani, P., 1984. Value and distribution in the classical economists and marx. Economic Policy. Palgrave Macmillan, Basingstoke and New York.
Oxf. Econ. Pap. 36 (June (2)), 291–325. Sraffa, P., 1960. Production of Commodities by Means of Commodities, ed.
Garegnani, P., 1992. Some notes for an analysis of accumulation. In: Halevi, J., Cambridge University Press, Cambridge.
Laibman, D., e Nell (a cura di), E.J. (Eds.), Beyond the Steady State: A Revival of Stirati, A., 2016. Real wages in the business cycle and the theory of income
Growth Theory. Macmillan, London, pp. 47–71. distribution: an unresolved conflict between theory and facts in mainstream
Garibaldi, P., Violante, G.L., 2005. The employment effects of severance payments macroeconomics. Camb. J. Econ. 40, 639–661.
with wage rigidities. Econ. J. 115, 799–832. Stockhammer, E., Onaran, Ö., Ederer, S., 2008. Functional income distribution and
Griffith, R., Harrison, R., Macartney, G., 2007. Product market reforms, labour aggregate demand in the Euro area. Camb. J. Econ. 33 (1), 139–159.
market institutions and unemployment. Econ. J. 117, 142–166. Stockhammer, E., 2013. Why Have Wage Shares Fallen? A Panel Analysis of the
Guerriero, M., Sen, K., 2012. What Determines the Share of Labour in National Determinants of Functional Income Distribution. ILO Working Paper.
Income? A Cross-country Analysis. IZA Discussion Paper No. 6643. Conditions of Work and Employment Series No. 35.
Heckman, J.J., 2007. Comments on ‘Are Protective Labor Market Institutions at the Storm, S., Naastepad, C.W.M., 2012. Macroeconomics Beyond the NAIRU, ed.
Root of Unemployment? A Critical Review of the Evidence’ by D. Howell, D., Harvard University Press, Cambridge, Massachusetts, and London, England.
Baker, A. Glyn and J. Schmitt, 2. Capitalism and Society (Issue 1). Togliatti, P., 1964. Capitalismo e riforme di struttura. Rinascita 28, 11, luglio.
Hicks, J.R., 1939. Value and Capital, ed. Oxford Clarendon Press, Oxford. Turrini, A., Koltay, G., Pierini, F., Goffard, C., Kiss, A., 2015. A decade of labour market
Howell, D.R., Rehm, M., 2009. Unemployment compensation and high European reforms in the EU: insights from the LABREF database. IZA J. Labor Policy 4, 12.
unemployment: a reassessment with new benefit indicators. Oxf. Rev. Econ. Vergeer, R., Kleinknecht, A., 2012. Do flexible labor markets indeed reduce
Policy 25 (1), 60–93. unemployment? A robustness check. Rev. Soc. Econ. 70, 451–467.
Howell, D.R., Baker, D., Glyn, A., Schmitt, J., 2007. Are protective labor market Williamson, S.D., 1986. Costly monitoring, financial intermediation, and
institutions really at the root of unemployment? A critical perspective on the equilibrium credit rationing. J. Monetary Econ. 18, 159–179.
statistical evidence. Capitalism Soc. 2 (1), 1–71. World Bank, 2013. World Development Report 2013: Jobs. World Bank
ILO, 2015. World Employment and Social Outlook: The Changing Nature of Jobs. Publications, Washington D.C.
International Labour Office, Geneva, Switzerland.
ILO, IMF, OECD, World Bank, 2015. Income inequality and labour income share in
G20 countries: trends, Impacts and Causes. In: Prepared for the G20 Labour